XGBoost模型与Logistic回归模型预测儿童TBTB合并MPP的价值  

Value of XGBoost model and Logistic regression model in predicting children with TBTB ambined with MPP

作　　者：刘小峰 洪智文 陈芳 严灿 谢齐放 LIU Xiaofeng;HONG Zhiwen;CHEN Fang;YAN Can;XIE Qifang(Department of Pediatrics,the Affiliated Changsha Central Hospital,Hengyang Medical School,University of South China,Changsha 410004,China)

机构地区：[1]南华大学衡阳医学院附属长沙中心医院儿科,长沙410004

出　　处：《华夏医学》2024年第5期55-61,共7页Acta Medicinae Sinica

基　　金：湖南省卫生健康委科研计划项目(D202306016238)。

摘　　要：目的比较XGBoost模型和Logistic回归模型在预测儿童气管支气管结核(TBTB)合并肺炎支原体肺炎(MPP)中的效能。方法收集2018年6月至2023年1月入院治疗的212例MPP患儿临床资料,根据是否合并TBTB将其分为合并组(n=42)与肺炎组(n=170)。通过XGBoost和Logistic回归构建儿童TBTB合并MPP的临床预测模型,采用受试者工作特征(ROC)曲线绘制两个模型在预测儿童TBTB合并MPP中的预测价值;采用Delong检验确定两个模型曲线下面积的差异。结果XGBoost模型筛选符合情况的相关变量为地区、咳嗽时间、发热史、咳嗽史、SIRI、WBC、湿罗音、SII,多因素Logistic回归分析发现,地区、发热史、咳嗽史、咳嗽时间、WBC是儿童TBTB合并MPP的独立危险因素(P<0.05),XGBoost曲线下面积为0.996,而Logistic回归模型曲线下面积仅为0.851。Delong检验发现,XGBoost模型预测效能高于Logistic回归模型,差异有统计学意义(P<0.01)。结论XGBoost模型优于Logistic回归模型,可作为诊断儿童TBTB合并MPP的有效工具。Objective To compare the predictive value of XGBoost model and Logistic regression model in children with tracheobronchial tuberculosis(TBTB)with mycoplasma pneumoniae pneumonia(MPP).Methods The clinical data of 212 children with MPP,admitted from June 2018 and January 2023,were collected.Based on whether the children were diagnosed with TBTB,they were divided into a combind group(n=42)and a pneumonia group(n=170).clinical prediction model for pediatric TBTB combined with MPP was constructed using XGBoost and Logistic regression.The receiver operating characteristic(ROC)curves was used to plot the predictive value of two models in predicting the of pediatric TBTB combined with MPP.The Delong test was used to determine the difference in the area under the curve between the two models.Results The relevant variables that matched the situation and filtered with the XGBoost model were region,cough duration,history of fever,history of cough,SIRI,white blood cell count(WBC),moist rale,and SII.Multivariate Logistic regression analysis revealed that region,fever history,cough history,cough duration,and WBC were risk factors for TBTB complicated with MPP in children(P<0.05).The area under the curve of the XGBoost model was 0.996,while the area under the curve of the Logistic regression model was only 0.851.Delong′s test found that the predictive efficacy of the XGBoost model was significantly higher than that of the Logistic model,with a statistically significant difference(P<0.01).Conclusion The XGBoost model is better than the Logistic regression model,and can be used as an effective tool for the diagnosis of TBTB combined with MPP in children.

关 键 词：气管支气管结核 肺炎支原体肺炎 XGBoost LOGISTIC回归 预测模型 

分 类 号：R725.6[医药卫生—儿科]